Analogous memory utilization-a necessary characteristic for large, adaptive neural net

When confronted with new knowledge structured in a familiar way humans often rely on analogy to make a one-to-one correspondence to the familiar structure. The author notes that efficient adaptive neural nets also have that characteristic. In fact, the absence of analogous memory utilization from a system will be obvious because the resulting uncontrolled redundancy will quickly fill new memory with an exponential explosion of redundant information. The resulting decrease in performance is rapid and dramatic. A discussion is presented of why proper use of available memory will produce analogous memory utilization automatically.<>